boost::thread_group and its associated join/interrupt mechanism has no equivalent in any C++XX standard.

The interrupt mechanism is basically a condition variable, per thread group, that interrupts boost::this_thread::sleep_for and boost::this_thread::sleep and raises boost::thread_interrupted in the member thread when group.interrupt() is called.

Luckily, it's only used for three

In e_os.h there is the following:

#    elif _MSC_VER<1300 && defined(I_CAN_LIVE_WITH_LNK4049)

Looking at https://en.wikipedia.org/wiki/Microsoft_Visual_C%2B%2B, it appears that this is for MSVC older than Visual Studio .NET 2002 version 7.0

I think this renaming of stdin/out/err and such should be removed because of the age, and because this is a completely undocumented compil

A couple of people have asked, so we should document how to migrate an Upspin store from one bucket to another (or indeed one storage backend to another).

This thread talks about gcp->gcp migration https://groups.google.com/d/msg/upspin/Lvx122keLJs/f09WhSu8BAAJ

Clients that are built around the API need some additional information to work effectively. This is the following:

1. Peer list API: show more information about connected peers including the block height / cumulative difficulty of the peer. Basically, so the API provides at least as much information as the server TUI does.

2. Status information: Similarly, the minimum amount of info the stat

Background

The binaries are built with most of the build flags already set, but when building from source you must specify which tags are desired. This is an important step to building a usable binary from source.

These flags should be listed in the docs, with what their use is for the entire system, and whether or not excluding it would break the normal function of the system.

A sect

Describe the solution you'd like
Improve Plan with refs to pointerPlan to not send more than 1 time the plan —> else return the same pointerPlan

This reduces the bandwidth usage and doesn't change the UX

There's no pure-JS OpenPGP (or crypto) library of similar quality.

The crypto really merits its own project, especially as many Node.js people are using rather poorly written alternatives.

See YahooArchive/end-to-end#32 for the parallel issue.

Problem: I cannot find a list of required packages. Following the build process using travis on ubuntu required at least python-pip, m4, zlib1g-dev and libnettle. A simple list for installing required packages would be nice.

Proposed Solution:
Something like what gnutls has: https://github.com/gnutls/gnutls (just search for apt-get)

We need rippled to improve its manifest support:

1. A rippled instance configured as a validator should offer an admin-only RPC command that can report the master public key, the ephemeral public key, the manifest sequence number and, if present, the domain associated with the manifest that the server is using. The command can be validator_info. If the server is not configured as a valida

Tendermint version (use tendermint version or git rev-parse --verify HEAD if installed from source):
github.com/tendermint/tendermint v0.32.5

ABCI app (name for bui

While experimenting within the Globaleaks application i just find out that the binding for SSL_CTX_set1_curves_list is currently not exposed.

This would be useful in order to define a preference on the order of the curves that should be used.

The feature would result particularly important especially for the fact that it seems that OpenSSL default order seems to be ```x25519, secp256r1, x448

https://scotthelme.co.uk/a-new-security-header-feature-policy/

I guess it will be easy to add, because we certainly don't use any of these features.
Or maybe "gyroscope" or so may be used by the random number generator in sjcl, but that's all.